Context-Aware and Locality-Constrained Coding for Image Categorization

Figure 2

The flowchart of the construction procedure of word-to-word cooccurrence matrix. (a) Extracting features using sift descriptor. (b) Assigning labels to the features using 𝑘-means. (c) Constructing word-to-word cooccurrence of each visual word. (d) Concatenating the cooccurrences of all visual words to form the final matrix.